Research Projects

This research addresses the critical need to understand and predict slope failures in the Indian Himalayan Region, a highly vulnerable area prone to landslides triggered by rainfall, earthquakes, and complex terrain conditions. The rationale is to fill gaps in current landslide hazard assessment by developing an integrated simulation framework that combines geospatial data, machine learning, laboratory experiments, and coupled multiphysics modelling. The scientific objective is to create a reliable, reproducible tool that can simulate slope stability under varying climatic and seismic conditions, improving early warning capabilities and risk mitigation. The key hypothesis is that integrating high-resolution terrain and geological data with machine learning–derived geotechnical parameters, constrained by laboratory thermo-hydro-mechanical testing, can enhance the accuracy of slope failure predictions. It further tests that coupled simulations incorporating rainfall, temperature, and seismic loading can realistically replicate observed landslide behaviors and deformation patterns. The main experiments begin with assembling multi-resolution digital elevation models and hydrometeorological data to reconstruct detailed terrain and environmental conditions. This is followed by training a convolutional neural network to predict spatial distributions of soil and rock mechanical properties using terrain features, vegetation indices, lithology, and soil texture. Soil and rock samples collected from key Himalayan sites will undergo cyclic triaxial tests to study behavior under repeated loading, temperature-controlled shear tests to assess freeze-thaw effects, and long-term creep tests to observe deformation under sustained stress. Hydraulic conductivity measurements under unsaturated conditions will also be conducted to model moisture flow. These laboratory results will be used to calibrate and constrain multiphysics simulations performed in an open-source framework, where coupled thermal, hydraulic, and mechanical processes under rainfall, temperature variations, and seismic shaking are modeled. Model outputs will be validated against satellite InSAR displacement data and historical landslide records through back-analysis and sensitivity studies. If successful, this research will significantly advance the fundamental understanding of how complex interactions between geology, climate, and seismic forces control slope stability. The framework will also provide practical applications by improving landslide susceptibility mapping, hazard assessment, and early warning systems tailored to Himalayan conditions. Ultimately, it offers a scalable, transparent approach to reduce landslide risks, safeguard communities, and guide sustainable land management in mountainous regions worldwide.